Improving Intraseasonal Prediction with a New Ensemble Generation Strategy
Improving Intraseasonal Prediction with a New Ensemble Generation Strategy
复制标题
DOI：
10.1175/mwr-d-13-00059.1
复制
发表时间：
2013-12-01
影响因子：
3.2
通讯作者：
Hendon, Harry H.
中科院分区：
文献类型：
--
作者：
Hudson, Debra;Marshall, Andrew G.;Hendon, Harry H.
澳大利亚气象局最近加强了对季节内气候变化进行耦合模式预测的能力。2013年3月之前运行的澳大利亚海洋大气预测模型（POAMA，第2版）季节预测预报系统（称为P2-S）不是为季节内预报设计的，在这方面存在缺陷。最值得注意的是，预报只在每个月的1日和15日初始化，在预报的前30天，集合传播的增长太慢，在季节内的时间尺度上是有用的。这些缺陷已经在系统升级中通过更频繁地初始化和通过增强系综生成来解决。新的集合生成方案是基于耦合繁殖的方法，并产生一个集合的扰动大气和海洋状态的初始化预报。与P2-S及其前身（1.5版）相比，该方案对澳大利亚降雨和温度的预报技巧产生了有利的影响。在POAMA-1.5中，系综是使用时滞大气初始条件但使用未扰动海洋初始条件产生的。P2-S使用了一个扰动海洋初始条件的集合，但只有一个单一的大气初始条件。使用耦合育种方法的预报性能的改善主要反映在第一个月预报的可靠性提高，但在预测季节内气候变率的重要驱动因素方面也有更高的技能，即Madden-Julian振荡和南部环形模式。结果说明了具有最佳集成生成策略的重要性。
The Australian Bureau of Meteorology has recently enhanced its capability to make coupled model forecasts of intraseasonal climate variations. The Predictive Ocean Atmosphere Model for Australia (POAMA, version 2) seasonal prediction forecast system in operations prior to March 2013, designated P2-S, was not designed for intraseasonal forecasting and has deficiencies in this regard. Most notably, the forecasts were only initialized on the 1st and 15th of each month, and the growth of the ensemble spread in the first 30 days of the forecasts was too slow to be useful on intraseasonal time scales. These deficiencies have been addressed in a system upgrade by initializing more often and through enhancements to the ensemble generation. The new ensemble generation scheme is based on a coupled-breeding approach and produces an ensemble of perturbed atmosphere and ocean states for initializing the forecasts. This scheme impacts favorably on the forecast skill of Australian rainfall and temperature compared to P2-S and its predecessor (version 1.5). In POAMA-1.5 the ensemble was produced using time-lagged atmospheric initial conditions but with unperturbed ocean initial conditions. P2-S used an ensemble of perturbed ocean initial conditions but only a single atmospheric initial condition. The improvement in forecast performance using the coupled-breeding approach is primarily reflected in improved reliability in the first month of the forecasts, but there is also higher skill in predicting important drivers of intraseasonal climate variability, namely the Madden-Julian oscillation and southern annular mode. The results illustrate the importance of having an optimal ensemble generation strategy.